Taiwan's Market Intelligence & Consulting Institute (MIC), a research arm of the state-backed Institute for Information Industry, released its 2026 IT services outlook on Tuesday, arguing that the proliferation of AI agents is forcing a fundamental break with the subscription-and-seat model that has defined software-as-a-service for two decades — and that companies slow to adapt face a potential reshuffling of their entire revenue structure.
MIC senior industry adviser Han Yang-ming (韓揚銘), who presented the findings at the MIC Forum Spring 2026 in Taipei, described the transition as "a DeepSeek Moment for the SaaS industry" — a sudden, structural shock rather than a gradual evolution. The core argument: when an AI agent can autonomously navigate multiple enterprise systems to complete a task end-to-end, companies stop paying for software access and start paying for results.
From operators to supervisors
Under the traditional SaaS model, humans operate platforms such as ERP or CRM systems, and vendors bill by the number of licensed accounts. In the emerging agent-driven model, that relationship inverts: AI agents handle cross-system process execution, while human workers shift into roles focused on setting objectives, defining operational boundaries, monitoring outputs, and bearing ultimate accountability for decisions. For IT service providers, Han said, the real competitive question is no longer whether a firm can deploy an AI model — it is whether the firm can architect human-machine systems that are trustworthy, controllable, and auditable.
Han also situated the report within a broader risk landscape, noting that geopolitical tension, political instability, and the threat of armed conflict have become the defining global risks of 2026. Information services, he argued, can no longer function merely as efficiency tools — they must become resilience infrastructure capable of sustaining enterprises through prolonged uncertainty.
Five global forces reshaping the market
MIC identified five broader dynamics converging on the global IT services landscape. AI is maturing beyond consumer productivity tools into general-purpose agents penetrating manufacturing, finance, and healthcare. Open-source models are gaining adoption due to cost and customisability advantages, driving hybrid deployment strategies that pair edge computing with centralised cloud infrastructure. Regulatory frameworks, meanwhile, are fragmenting: the EU is prioritising rights protection, the US is focused on innovation competitiveness, and China is framing AI governance around national security — creating an uneven compliance environment for firms operating across borders.
Two additional pressures complete MIC's global picture. Agent governance is becoming a frontline security issue: as AI moves from generating content to executing transactions and mobilising enterprise systems, cybersecurity can no longer function as an optional layer. Blockchain, MIC projected, is also transitioning from proof-of-concept experimentation toward cross-industry standardisation, positioning it as foundational trust infrastructure for the digital economy.
Six strategies for Taiwan's IT sector
Against this backdrop, MIC proposed six strategic directions for Taiwan-based IT service providers. Han said firms must first build proprietary data moats, since domain-specific knowledge, process data, and customer context will be the primary sources of differentiation as model capabilities commoditise. Second, regulatory compliance must be embedded in product architecture from the outset — particularly in finance, healthcare, and public services — rather than treated as an external audit requirement.
Third, human oversight must be designed into critical process nodes as an engineering requirement, not a talking point. Fourth, providers should join ecosystems and build platform leverage through partner networks, generating network effects that isolated competitors cannot quickly replicate. Fifth, billing models should gradually shift toward outcome-based structures — charging per completed transaction, resolved service request, or measurable business result — running in parallel with existing seat-based fees. Sixth, Taiwan's IT sector must develop a new generation of engineers fluent in human-AI collaborative system design, a capability MIC described as both the scarcest and the most critical in the industry.
The "HI+AI CODE" framework
MIC distilled these priorities into a framework it calls "HI+AI CODE," built around four pillars: Human in the Loop, Compliance, Outcome Data, and Ecosystem. The institute framed it not as a technology checklist but as a description of a role transformation — from tool supplier to enterprise operating architecture designer. As AI agents increasingly automate process execution, Han said, what IT service companies will ultimately sell is not software functionality, but measurable outcomes, trustworthy governance, and systems built to work alongside human workers. （Related： Taiwan’s 2026 ICT Outlook: How AI Infrastructure Is Reshaping the Global Supply Chain ｜ Latest ）
